Today, social behavior and transaction data are seldom correlated to fully understand the social graph of customers that can lead to business strategy decisions like special handling of key influencers. While social media offers some insights into key influencers, such correlation with transactional data and behavioral data would take that to the next level.
Read this discussion with Dmitri Williams to understand where we’re going.
Sramana Mitra: Let’s introduce our audience to you and give some context about what vantage point you’re looking at the Big Data world from.
Dmitri Williams: I wear a couple of different hats and I have a couple of titles. One is Associate Professor at the University of Southern California where I teach at the Annenberg School of Communication. I teach in the program for online communities. The other is as the CEO of Ninja Metrics, which is a large-scale data analytics company – largely B2B and serving people in the gaming industry. We are expanding to some other spaces too.
Sramana Mitra: Tell me more about the research work that you do in social sciences and how does Big Data apply to your work.
Dmitri Williams: When I first started, I didn’t think about it at all. As social scientists, we normally start off with what we thought were large data sets. We would get survey data from 10,000 people or experiment data from a few hundred people. That was big for us. About 10 years ago, I started getting involved with videogame companies. It was obvious that they had these very large treasure troves of behavioral data, which were of much higher quality than the self-reported surveys. There are millions of interactions and transactions. I convinced Sony Online Entertainment to let us do a research study and they sent me a three terabyte drive and I had no idea what to do with it. I very quickly made friends with computer scientists. I started learning what the big side of Big Data meant.
The past 10 years or so has seen us forming research teams and figuring the right kind of processes to marry this chocolate and peanut butter like elements of social science and computer science. It has been a very interesting journey and one with a lot of stumbles, failures, and successes along the way. It’s very messy when you do something interdisciplinary. Today, we have a large team both on the commercial side and the academic side that blends the work teams of social scientists and computer scientists tackling specific problems. We’ve now got a pretty strong methodology that we’ve built up through a lot of projects that have been funded, at first, by government agencies like intelligence communities and the army. Now, my company is also funded on the commercial side.